
# Importing packages - Pandas, Numpy, Seaborn, Scipy
import pandas as pd, numpy as np, matplotlib.pyplot as plt, seaborn as sns, sys
import matplotlib.style as style; style.use('fivethirtyeight')
from scipy.stats import zscore, norm
from sklearn.preprocessing import LabelEncoder

# Importing packages for KNN(supervised learning)
from sklearn.preprocessing import StandardScaler
from sklearn.metrics import classification_report, confusion_matrix, roc_auc_score, roc_curve, accuracy_score
from sklearn.model_selection import train_test_split, GridSearchCV, StratifiedKFold
from sklearn.neighbors import KNeighborsClassifier

# Suppress warnings
import warnings; warnings.filterwarnings('ignore')
pd.options.display.max_rows = 4000


# importing dataset onto python
# 1st set
Normal = pd.read_csv('Part1 - Normal.csv')
Normal.head()


Normal.shape

(100, 7)


#2nd dataset
typeH = pd.read_csv('Part1 - Type_H.csv')
typeH.head()


typeH.shape

(60, 7)


# 3rd dataset
typeS = pd.read_csv('Part1 - Type_S.csv')
typeS.head()


typeS.shape

(150, 7)


# Merging all 3 datasets onto one.
Final_set = [Normal,typeH,typeS]
final_set = pd.concat(Final_set)


final_set.shape
#100+60+150

(310, 7)


final_set.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 310 entries, 0 to 149
Data columns (total 7 columns):
 #   Column       Non-Null Count  Dtype  
---  ------       --------------  -----  
 0   P_incidence  310 non-null    float64
 1   P_tilt       310 non-null    float64
 2   L_angle      310 non-null    float64
 3   S_slope      310 non-null    float64
 4   P_radius     310 non-null    float64
 5   S_Degree     310 non-null    float64
 6   Class        310 non-null    object 
dtypes: float64(6), object(1)
memory usage: 19.4+ KB


final_set.head(10)


final_set.tail(10)


final_set.Class.value_counts()

Type_S    133
Normal     73
Type_H     37
Nrmal      27
type_h     23
tp_s       17
Name: Class, dtype: int64


final_set.isnull().value_counts()

P_incidence  P_tilt  L_angle  S_slope  P_radius  S_Degree  Class
False        False   False    False    False     False     False    310
dtype: int64


# correcting CLASS column
final_set = final_set.replace(to_replace='Nrmal',value='Normal',regex=True)
final_set = final_set.replace(to_replace='type_h',value='Type_H',regex=True)
final_set = final_set.replace(to_replace='tp_s',value='Type_S',regex=True)


final_set.Class.value_counts().to_frame()


final_set.all().to_frame()


final_set.Class.value_counts(normalize=True)*100

Type_S    48.387097
Normal    32.258065
Type_H    19.354839
Name: Class, dtype: float64


final_set.reset_index(inplace=True)
final_set


f = ['index']
final_set.drop(columns=f, inplace=True)
final_set


# checking the data in random order.
final_set.sample(10)


final_set.describe().T


# 1. We'll draw the pairplot to check and compare each variable and try and find relations among them.
sns.pairplot(final_set)

<seaborn.axisgrid.PairGrid at 0x7faaedb8cf40>


# analysing categorical variable.
import plotly.express as px
fig = px.histogram(final_set.Class)
fig.show()


# analysing each variable one by one:

P_incidence_mean = final_set.P_incidence.mean()

fig, ax_hist = plt.subplots(figsize = (12.8, 6))
sns.distplot(final_set.P_incidence)
ax_hist.axvline(P_incidence_mean, color = 'r', linestyle = '--', label = 'Mean')

<matplotlib.lines.Line2D at 0x7faaef7f0370>


P_radious_mean = final_set.P_radius.mean()

fig, ax_hist = plt.subplots(figsize = (10, 5))
sns.distplot(final_set.P_radius)
ax_hist.axvline(P_radious_mean, color = 'r', linestyle = '--', label = 'Mean')

<matplotlib.lines.Line2D at 0x7fab02e5ec10>


P_tilt_mean = final_set.P_tilt.mean()

fig, ax_hist = plt.subplots(figsize = (10,5))
sns.distplot(final_set.P_tilt)
ax_hist.axvline(P_tilt_mean, color = 'r', linestyle = '--', label = 'Mean')

<matplotlib.lines.Line2D at 0x7fab02ef4d30>


S_slope_mean = final_set.S_slope.mean()

fig, ax_hist = plt.subplots(figsize = (10,5))
sns.distplot(final_set.S_slope)
ax_hist.axvline(S_slope_mean, color = 'r', linestyle = '--', label = 'Mean')

<matplotlib.lines.Line2D at 0x7fab03a9e190>


L_angle_mean = final_set.L_angle.mean()

fig, ax_hist = plt.subplots(figsize = (10,5))
sns.distplot(final_set.L_angle)
ax_hist.axvline(L_angle_mean, color = 'r', linestyle = '--', label = 'Mean')

<matplotlib.lines.Line2D at 0x7faae6d19ac0>


S_Degree_mean = final_set.S_Degree.mean()

fig, ax_hist = plt.subplots(figsize = (10,5))
sns.distplot(final_set.S_Degree)
ax_hist.axvline(S_Degree_mean, color = 'r', linestyle = '--', label = 'Mean')

<matplotlib.lines.Line2D at 0x7fab02ebbd90>


fig = px.scatter(x = final_set.P_incidence, y = final_set.S_Degree, color=final_set.Class)
fig.show()


s = sns.boxplot(x=final_set['P_incidence'],y=final_set['Class'])


final_set[final_set['P_incidence'] > 100]


s = sns.boxplot(x=final_set['S_Degree'],y=final_set['Class'])


final_set[final_set['S_Degree'] > 300]


s = sns.boxplot(x=final_set['S_slope'],y=final_set['Class'])


final_set[final_set['S_slope'] > 120]


s = sns.boxplot(x=final_set['P_tilt'],y=final_set['Class'])


s = sns.boxplot(x=final_set['P_radius'],y=final_set['Class'])


final_set[final_set['P_radius'] > 150]


s = sns.boxplot(x=final_set['L_angle'],y=final_set['Class'])


final_set[final_set['L_angle'] > 100]


final_set.drop(index=215, inplace=True)
final_set.count()

P_incidence    309
P_tilt         309
L_angle        309
S_slope        309
P_radius       309
S_Degree       309
Class          309
dtype: int64


correlation = final_set.corr()
plt.figure(figsize = (15, 10))
plt.xticks(fontsize = 15, rotation = 45)
plt.yticks(fontsize = 15)
plt.title('Correlation HeatMap', fontsize = 15)
sns.heatmap(correlation, annot = True, cmap = 'YlGnBu')

<AxesSubplot:title={'center':'Correlation HeatMap'}>


# dropping P-Radious  ---> done after regular modelling below


final_set.dtypes

P_incidence    float64
P_tilt         float64
L_angle        float64
S_slope        float64
P_radius       float64
S_Degree       float64
Class           object
dtype: object


final_set['Class'] = final_set['Class'].astype('category')


final_set.dtypes

P_incidence     float64
P_tilt          float64
L_angle         float64
S_slope         float64
P_radius        float64
S_Degree        float64
Class          category
dtype: object


#Copying the dataset for encoding.
final_set_encoded = final_set.copy(deep = True)
final_set_encoded.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 309 entries, 0 to 309
Data columns (total 7 columns):
 #   Column       Non-Null Count  Dtype   
---  ------       --------------  -----   
 0   P_incidence  309 non-null    float64 
 1   P_tilt       309 non-null    float64 
 2   L_angle      309 non-null    float64 
 3   S_slope      309 non-null    float64 
 4   P_radius     309 non-null    float64 
 5   S_Degree     309 non-null    float64 
 6   Class        309 non-null    category
dtypes: category(1), float64(6)
memory usage: 27.3 KB


enc = LabelEncoder()
final_set_encoded['Class'] = enc.fit_transform(final_set_encoded['Class']).astype(int)


final_set_encoded.dtypes

P_incidence    float64
P_tilt         float64
L_angle        float64
S_slope        float64
P_radius       float64
S_Degree       float64
Class            int64
dtype: object


# dividing data into Predicator(X) vs Target(y)-

X = final_set_encoded.drop('Class',axis = 1)
print('Sample Predicator data:\n',X.head(5))
y = final_set_encoded['Class']
print('\nSample Traget data:\n',y.head(5))

Sample Predicator data:
    P_incidence     P_tilt    L_angle    S_slope    P_radius  S_Degree
0    38.505273  16.964297  35.112814  21.540976  127.632875  7.986683
1    54.920858  18.968430  51.601455  35.952428  125.846646  2.001642
2    44.362490   8.945435  46.902096  35.417055  129.220682  4.994195
3    48.318931  17.452121  48.000000  30.866809  128.980308 -0.910941
4    45.701789  10.659859  42.577846  35.041929  130.178314 -3.388910

Sample Traget data:
 0    0
1    0
2    0
3    0
4    0
Name: Class, dtype: int64


X.corrwith(y).plot.bar(grid = True)
plt.title('Correlation with Class', fontsize = 20)

Text(0.5, 1.0, 'Correlation with Class')


X_std = pd.DataFrame(StandardScaler().fit_transform(X))
X_std.columns = X.columns #the column names are passed to the standardized dataframe
X_std.head()


X_std.describe().T


# Splitting data in train and test 

X_train, X_test, y_train, y_test = train_test_split(X_std, y, test_size = 0.3, random_state = 10)


X_train.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 216 entries, 210 to 265
Data columns (total 6 columns):
 #   Column       Non-Null Count  Dtype  
---  ------       --------------  -----  
 0   P_incidence  216 non-null    float64
 1   P_tilt       216 non-null    float64
 2   L_angle      216 non-null    float64
 3   S_slope      216 non-null    float64
 4   P_radius     216 non-null    float64
 5   S_Degree     216 non-null    float64
dtypes: float64(6)
memory usage: 11.8 KB


y_train.value_counts()

2    100
0     71
1     45
Name: Class, dtype: int64


y_test.value_counts()

2    49
0    29
1    15
Name: Class, dtype: int64


# check for the Target balancing
print('\nTarget balancing between train and test data: \n')
print('Ratio of Class = 0: ', (y_test == 0).count() / (y_train == 0).count())
print('Ratio of Class = 1: ', (y_test == 1).count() / (y_train == 1).count())
print('Ratio of Class = 2: ', (y_test == 2).count() / (y_train == 2).count())

Target balancing between train and test data: 

Ratio of Class = 0:  0.4305555555555556
Ratio of Class = 1:  0.4305555555555556
Ratio of Class = 2:  0.4305555555555556


# checking shape of train and test data:
a, b = X_train.shape
print('The training set comprises of', a, 'rows and', b, 'columns.')

c, d = X_test.shape
print('The testing set comprises of', c, 'rows and', d, 'columns.')

The training set comprises of 216 rows and 6 columns.
The testing set comprises of 93 rows and 6 columns.


# We'll start with basic KNN model with n=3 and then later, we'll find the best neighbour and model accuracies.

NHH1 = KNeighborsClassifier(n_neighbors=3, weights='distance')


NHH1.fit(X_train,y_train)

KNeighborsClassifier(n_neighbors=3, weights='distance')


# checking predictions
predicted_lab = NHH1.predict(X_test)
NHH1.score(X_test,y_test)

0.7634408602150538


from sklearn.metrics import confusion_matrix
from sklearn.neighbors import KNeighborsClassifier
y_pred = NHH1.predict(X_test)
confusion_matrix(y_test, y_pred)

array([[17, 11,  1],
       [ 4, 11,  0],
       [ 4,  2, 43]])


# Now, we'll be automating the task to find the best fit Model and neighbour for the KNN model.

# odd list for Neighbour range:
myList = list(range(1, 30))

# selecting only odd values to pass:
neighbors = list(filter(lambda x: x % 2 != 0, myList))
accuracy_scores = [] # empty list to hold thre result.


for k in neighbors:
    knn = KNeighborsClassifier(n_neighbors = k)
    knn.fit(X_train, y_train)
    # predict the response
    y_predictKNN = knn.predict(X_test)
    # evaluate accuracy
    scores = accuracy_score(y_test, y_predictKNN)
    accuracy_scores.append(scores)

# changing to misclassification error
MSE = [1 - x for x in accuracy_scores]

# Best k value: we found the best value of 'k' with the help of Misclassification error.
optimal_k = neighbors[MSE.index(min(MSE))]
print('The best value of neighbors is: ', optimal_k)

The best value of neighbors is:  5


#Using optimal_k from above as the final model for prediction
knn = KNeighborsClassifier(n_neighbors = optimal_k)

#Fit on train data
knn.fit(X_train, y_train)

#fit on test data
y_predictKNN = knn.predict(X_test)

# evaluate accuracy
print('\nAccuracy Score of Training Data: ', knn.score(X_train, y_train))
modelKNN_score = accuracy_score(y_test,y_predictKNN)

cfm_KNN = confusion_matrix(y_test, y_predictKNN)
print('Accuracy Score of Test Data: ', modelKNN_score)

print('\nClassification Report of KNN Model:\n ', classification_report(y_test, y_predictKNN))
print('\nConfusion Matrix for our Model:\n', cfm_KNN)

#visualization of confusion matrix in the form of a heatmap
plt.figure(figsize = (12, 8))
sns.heatmap(cfm_KNN, annot = True, cmap = 'YlGnBu', fmt = 'd')
plt.xlabel('Actual Classes', fontsize = 15)
plt.ylabel('Predicted Classes', fontsize = 15)
plt.title('Confusion Matrix HeatMap', fontsize = 15)

Accuracy Score of Training Data:  0.8796296296296297
Accuracy Score of Test Data:  0.8064516129032258

Classification Report of KNN Model:
                precision    recall  f1-score   support

           0       0.72      0.72      0.72        29
           1       0.52      0.73      0.61        15
           2       1.00      0.88      0.93        49

    accuracy                           0.81        93
   macro avg       0.75      0.78      0.76        93
weighted avg       0.84      0.81      0.82        93


Confusion Matrix for our Model:
 [[21  8  0]
 [ 4 11  0]
 [ 4  2 43]]

Text(0.5, 1.0, 'Confusion Matrix HeatMap')


from sklearn.metrics import precision_recall_fscore_support, mean_absolute_error
precision_KNN, recall_KNN, f1_score_KNN, support = precision_recall_fscore_support(y_test, y_predictKNN, average = 'macro')
print('Precision Score :', '%0.2f' % precision_KNN)
print('Recall Score :', '%0.2f' % recall_KNN)
print('F1-Score :', '%0.2f' % f1_score_KNN)
print('Accuracy Score :','%0.2f' % modelKNN_score)

Precision Score : 0.75
Recall Score : 0.78
F1-Score : 0.76
Accuracy Score : 0.81


f1 = final_set_encoded.copy(deep=True)


f1.drop(columns='P_radius',inplace=True)
f1.head(5)


# splitting the data as A = X and B = y
A = f1.drop('Class',axis = 1)
B = f1['Class']


# normalizing the data with z-score
f1 = f1.apply(zscore)


f1.describe().T


# splitting data for train and test
X_train, X_test, y_train, y_test = train_test_split(A, B, test_size = 0.3, random_state = 10)


# fitting the KNN model for k=5 (proven best fit above)
NHH1 = KNeighborsClassifier(n_neighbors=5, weights='distance')
NHH1.fit(X_train,y_train)

KNeighborsClassifier(weights='distance')


predicted_lab = NHH1.predict(X_test)
NHH1.score(X_test,y_test)

0.8279569892473119


precision_KNN, recall_KNN, f1_score_KNN, support = precision_recall_fscore_support(y_test, predicted_lab, average = 'macro')
print('Precision Score :', '%0.2f' % precision_KNN)
print('Recall Score :', '%0.2f' % recall_KNN)
print('F1-Score :', '%0.2f' % f1_score_KNN)
print('Accuracy Score :','%0.2f' % modelKNN_score)

Precision Score : 0.78
Recall Score : 0.81
F1-Score : 0.78
Accuracy Score : 0.81


# We have already imported all the necessary libraries above, we'll import more when needed
# importing data from file 1
Part2_1 = pd.read_csv('Part2 - Data1.csv')
Part2_1.head(5)


# importing data from file 2
Part2_2 = pd.read_csv('Part2 -Data2.csv')
Part2_2.head(5)


Part2_1.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 5000 entries, 0 to 4999
Data columns (total 8 columns):
 #   Column               Non-Null Count  Dtype  
---  ------               --------------  -----  
 0   ID                   5000 non-null   int64  
 1   Age                  5000 non-null   int64  
 2   CustomerSince        5000 non-null   int64  
 3   HighestSpend         5000 non-null   int64  
 4   ZipCode              5000 non-null   int64  
 5   HiddenScore          5000 non-null   int64  
 6   MonthlyAverageSpend  5000 non-null   float64
 7   Level                5000 non-null   int64  
dtypes: float64(1), int64(7)
memory usage: 312.6 KB


Part2_2.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 5000 entries, 0 to 4999
Data columns (total 7 columns):
 #   Column               Non-Null Count  Dtype  
---  ------               --------------  -----  
 0   ID                   5000 non-null   int64  
 1   Mortgage             5000 non-null   int64  
 2   Security             5000 non-null   int64  
 3   FixedDepositAccount  5000 non-null   int64  
 4   InternetBanking      5000 non-null   int64  
 5   CreditCard           5000 non-null   int64  
 6   LoanOnCard           4980 non-null   float64
dtypes: float64(1), int64(6)
memory usage: 273.6 KB


print('Shape of dataset 1 :',Part2_1.shape)
print('Shape of dataset 2 :',Part2_2.shape)

Shape of dataset 1 : (5000, 8)
Shape of dataset 2 : (5000, 7)


# mergign of the dataset.

Data = pd.merge(Part2_1,Part2_2,on='ID')


Data.head(10)


Data.shape

(5000, 14)


Data.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 5000 entries, 0 to 4999
Data columns (total 14 columns):
 #   Column               Non-Null Count  Dtype  
---  ------               --------------  -----  
 0   ID                   5000 non-null   int64  
 1   Age                  5000 non-null   int64  
 2   CustomerSince        5000 non-null   int64  
 3   HighestSpend         5000 non-null   int64  
 4   ZipCode              5000 non-null   int64  
 5   HiddenScore          5000 non-null   int64  
 6   MonthlyAverageSpend  5000 non-null   float64
 7   Level                5000 non-null   int64  
 8   Mortgage             5000 non-null   int64  
 9   Security             5000 non-null   int64  
 10  FixedDepositAccount  5000 non-null   int64  
 11  InternetBanking      5000 non-null   int64  
 12  CreditCard           5000 non-null   int64  
 13  LoanOnCard           4980 non-null   float64
dtypes: float64(2), int64(12)
memory usage: 585.9 KB


Data.isnull().value_counts().to_frame().T


Data.describe(include='all').T


Data.HiddenScore.value_counts().to_frame()


Data.Level.value_counts().to_frame()


Data.MonthlyAverageSpend.unique().sum()

463.07


Data.Security.value_counts().to_frame()


Data.FixedDepositAccount.value_counts().to_frame()


Data.InternetBanking.value_counts().to_frame()


Data.CreditCard.value_counts().to_frame()


Data.LoanOnCard.value_counts().to_frame()


Data[Data.LoanOnCard.isnull()]


# we see here that from 5000 entries, we have 20 entries as null for one of the target variables. 
#Since the number of null records are less, we'll drop these 20 records.

Data.dropna(inplace=True)


Data.shape

(4980, 14)


Data.isnull().sum()

ID                     0
Age                    0
CustomerSince          0
HighestSpend           0
ZipCode                0
HiddenScore            0
MonthlyAverageSpend    0
Level                  0
Mortgage               0
Security               0
FixedDepositAccount    0
InternetBanking        0
CreditCard             0
LoanOnCard             0
dtype: int64


Data.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 4980 entries, 9 to 4999
Data columns (total 14 columns):
 #   Column               Non-Null Count  Dtype  
---  ------               --------------  -----  
 0   ID                   4980 non-null   int64  
 1   Age                  4980 non-null   int64  
 2   CustomerSince        4980 non-null   int64  
 3   HighestSpend         4980 non-null   int64  
 4   ZipCode              4980 non-null   int64  
 5   HiddenScore          4980 non-null   int64  
 6   MonthlyAverageSpend  4980 non-null   float64
 7   Level                4980 non-null   int64  
 8   Mortgage             4980 non-null   int64  
 9   Security             4980 non-null   int64  
 10  FixedDepositAccount  4980 non-null   int64  
 11  InternetBanking      4980 non-null   int64  
 12  CreditCard           4980 non-null   int64  
 13  LoanOnCard           4980 non-null   float64
dtypes: float64(2), int64(12)
memory usage: 583.6 KB


# we'll now make changes on the variable types.
# We have already noted all the categorical variables above, we'll make the change on datatypes now.


var = ['HiddenScore',
      'Level',
      'Security',
      'FixedDepositAccount',
      'InternetBanking',
      'CreditCard',
      'LoanOnCard']
for v in var:
    Data[v] = Data[v].astype('category')
    
Data.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 4980 entries, 9 to 4999
Data columns (total 14 columns):
 #   Column               Non-Null Count  Dtype   
---  ------               --------------  -----   
 0   ID                   4980 non-null   int64   
 1   Age                  4980 non-null   int64   
 2   CustomerSince        4980 non-null   int64   
 3   HighestSpend         4980 non-null   int64   
 4   ZipCode              4980 non-null   int64   
 5   HiddenScore          4980 non-null   category
 6   MonthlyAverageSpend  4980 non-null   float64 
 7   Level                4980 non-null   category
 8   Mortgage             4980 non-null   int64   
 9   Security             4980 non-null   category
 10  FixedDepositAccount  4980 non-null   category
 11  InternetBanking      4980 non-null   category
 12  CreditCard           4980 non-null   category
 13  LoanOnCard           4980 non-null   category
dtypes: category(7), float64(1), int64(6)
memory usage: 346.1 KB


# we'll start with Basic EDA:
Data.describe().T


# we'll now start with Uni, Bi and multivariate analysis:

#1. Age

sns.distplot(Data.Age,color='green').set_title('Desnsity plot for Age')
plt.show()
sns.boxplot(Data.Age, color='blue').set_title('Box plot for Age')
plt.show()


#2. CustomerSince

sns.distplot(Data.CustomerSince,color='Orange').set_title('Desnsity plot for CustomerSince')
plt.show()
sns.boxplot(Data.CustomerSince, color='brown').set_title('Box plot for CustomerSince')
plt.show()


#3. HighestSpend

sns.distplot(Data.HighestSpend,color='pink').set_title('Desnsity plot for Highest Spent')
plt.show()
plt.figure(figsize = (15, 8))
sns.boxplot(Data.HighestSpend, color='gold').set_title('Box plot for Highest Spent')
plt.show()


# We can see, that there are outliers present here. We'll find the number of outliers.
q1 = np.percentile(Data.HighestSpend,25)
q3 = np.percentile(Data.HighestSpend,75)
iqr = q3-q1
u_f = q3 + 1.5*iqr
print('\nMax wisker is at:',u_f)
print("\nNo. of outiers:",Data.HighestSpend [Data['HighestSpend'] > u_f].count())

Max wisker is at: 186.5

No. of outiers: 96


# We'll not be conducting action on Zipcode as id does not provide much info. --> will drop it.

# 4. Monthly Avg. Spent

sns.distplot(Data.MonthlyAverageSpend,color='green').set_title('Desnsity plot for MonthlyAvgSpent')
plt.show()
plt.figure(figsize = (15, 8))
sns.boxplot(Data.MonthlyAverageSpend, color='purple').set_title('Box plot for MonthlyAvgSpent')
plt.show()


# We can see, that there are outliers present here. We'll find the number of outliers.
q1 = np.percentile(Data.MonthlyAverageSpend,25)
q3 = np.percentile(Data.MonthlyAverageSpend,75)
iqr = q3-q1
u_f = q3 + 1.5*iqr
print('\nMax wisker is at:',u_f)
print("\nNo. of outiers:",Data.MonthlyAverageSpend [Data['MonthlyAverageSpend'] > u_f].count())

Max wisker is at: 5.262499999999999

No. of outiers: 324


# 4. Morgage

plt.figure(figsize = (15, 6))
sns.distplot(Data.Mortgage,color='teal').set_title('Desnsity plot for Morgage')
plt.show()
plt.figure(figsize = (15, 6))
sns.boxplot(Data.Mortgage, color='yellow').set_title('Box plot for Morgage')
plt.show()


# We can see, that there are outliers present here. We'll find the number of outliers.
q1 = np.percentile(Data.Mortgage,25)
q3 = np.percentile(Data.Mortgage,75)
iqr = q3-q1
u_f = q3 + 1.5*iqr
print('\nMax wisker is at:',u_f)
print("\nNo. of outiers:",Data.Mortgage [Data['Mortgage'] > u_f].count())

Max wisker is at: 252.5

No. of outiers: 291


Data.Mortgage.value_counts().head(5)

0     3447
98      17
91      16
83      16
89      16
Name: Mortgage, dtype: int64


# Some analysis on Categorical data:


Data.LoanOnCard.value_counts().plot.pie(autopct='%1.1f%%',shadow=True)

<AxesSubplot:ylabel='LoanOnCard'>


Data.HiddenScore.value_counts().plot.pie(autopct='%1.1f%%',shadow=True)

<AxesSubplot:ylabel='HiddenScore'>


Data.FixedDepositAccount.value_counts().plot.pie(autopct='%1.1f%%',shadow=True)

<AxesSubplot:ylabel='FixedDepositAccount'>


Data.CreditCard.value_counts().plot.pie(autopct='%1.1f%%',shadow=True)

<AxesSubplot:ylabel='CreditCard'>


Data.Level.value_counts().plot.pie(autopct='%1.1f%%',shadow=True)

<AxesSubplot:ylabel='Level'>


# we'll have a check on CategoricalvsCategorical as well now:

plt.figure(figsize = (10, 6))
sns.countplot(x=Data.HiddenScore,hue=Data.LoanOnCard)

<AxesSubplot:xlabel='HiddenScore', ylabel='count'>


plt.figure(figsize = (10, 6))
sns.countplot(x=Data.CreditCard,hue=Data.LoanOnCard)

<AxesSubplot:xlabel='CreditCard', ylabel='count'>


plt.figure(figsize = (10, 6))
sns.countplot(x=Data.Security,hue=Data.LoanOnCard)

<AxesSubplot:xlabel='Security', ylabel='count'>


plt.figure(figsize = (10, 6))
sns.countplot(x=Data.Security,hue=Data.CreditCard)

<AxesSubplot:xlabel='Security', ylabel='count'>


correlation = Data.corr()
plt.figure(figsize = (15, 10))
plt.xticks(fontsize = 15, rotation = 45)
plt.yticks(fontsize = 15)
plt.title('Correlation HeatMap', fontsize = 15)
sns.heatmap(correlation, annot = True, cmap = 'YlGnBu')

<AxesSubplot:title={'center':'Correlation HeatMap'}>


# from above analysis, EDA and Corr, we found that columns Age, Customer Since and ZIpcode are not much of contributers into our dataset.
# hence, we'll drop these columns for model building.

Data.drop(['Age','CustomerSince','ZipCode'], axis = 1, inplace=True)


Data.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 4980 entries, 9 to 4999
Data columns (total 11 columns):
 #   Column               Non-Null Count  Dtype   
---  ------               --------------  -----   
 0   ID                   4980 non-null   int64   
 1   HighestSpend         4980 non-null   int64   
 2   HiddenScore          4980 non-null   category
 3   MonthlyAverageSpend  4980 non-null   float64 
 4   Level                4980 non-null   category
 5   Mortgage             4980 non-null   int64   
 6   Security             4980 non-null   category
 7   FixedDepositAccount  4980 non-null   category
 8   InternetBanking      4980 non-null   category
 9   CreditCard           4980 non-null   category
 10  LoanOnCard           4980 non-null   category
dtypes: category(7), float64(1), int64(3)
memory usage: 389.3 KB


Data.drop('ID',axis=1,inplace=True)


Data.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 4980 entries, 9 to 4999
Data columns (total 10 columns):
 #   Column               Non-Null Count  Dtype   
---  ------               --------------  -----   
 0   HighestSpend         4980 non-null   int64   
 1   HiddenScore          4980 non-null   category
 2   MonthlyAverageSpend  4980 non-null   float64 
 3   Level                4980 non-null   category
 4   Mortgage             4980 non-null   int64   
 5   Security             4980 non-null   category
 6   FixedDepositAccount  4980 non-null   category
 7   InternetBanking      4980 non-null   category
 8   CreditCard           4980 non-null   category
 9   LoanOnCard           4980 non-null   category
dtypes: category(7), float64(1), int64(2)
memory usage: 350.4 KB


# Also, after making checks, we can confirm that out of all categorical values,
# Hidden Score, Level, Fixed Deposit and LoanOnCard are more contributing than Security, InternetBanking and CreditCard.

# Hence, we'll driop these values as well:

Data.drop(['CreditCard','InternetBanking','Security'],axis=1,inplace=True)
Data.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 4980 entries, 9 to 4999
Data columns (total 7 columns):
 #   Column               Non-Null Count  Dtype   
---  ------               --------------  -----   
 0   HighestSpend         4980 non-null   int64   
 1   HiddenScore          4980 non-null   category
 2   MonthlyAverageSpend  4980 non-null   float64 
 3   Level                4980 non-null   category
 4   Mortgage             4980 non-null   int64   
 5   FixedDepositAccount  4980 non-null   category
 6   LoanOnCard           4980 non-null   category
dtypes: category(4), float64(1), int64(2)
memory usage: 335.6 KB


# We'll now fix the outliers and replace them with the mean value for more accurate model developemnt.

# variables with outliers:
col=['HighestSpend','MonthlyAverageSpend','Mortgage']

# magic hapenning below :P
#we'll impute the values of Mean for all the outliers present in 3 variabless

for c in col:
    #getting upper lower quartile values
    q25,q75=np.percentile(Data[c],25),np.percentile(Data[c],75)
    IQR=q75-q25
    Threshold=IQR*1.5
    lower,upper=q25-Threshold,q75+Threshold
    Outliers=[i for i in Data[c] if i < lower or i > upper]
    print('{}Number of outliers in {} Before Imputing : {}'.format('\033[1m',c,len(Outliers)))
    #taking mean of a column without considering outliers
    Data_include = Data.loc[(Data[c] >= lower) & (Data[c] <= upper)]
    mean=int(Data_include[c].mean())
    print('{}Mean {} is {}'.format('\033[1m',c,mean))
    #imputing outliers with mean
    Data[c]=np.where(Data[c]>upper,mean,Data[c])
    Data[c]=np.where(Data[c]<lower,mean,Data[c])
    Outliers=[i for i in Data[c] if i < lower or i > upper]
    print('{}Number of outliers in {} After Imputing : {}'.format('\033[1m',c,len(Outliers)))  
    print('\n')

Number of outliers in HighestSpend Before Imputing : 96
Mean HighestSpend is 71
Number of outliers in HighestSpend After Imputing : 0


Number of outliers in MonthlyAverageSpend Before Imputing : 324
Mean MonthlyAverageSpend is 1
Number of outliers in MonthlyAverageSpend After Imputing : 0


Number of outliers in Mortgage Before Imputing : 291
Mean Mortgage is 38
Number of outliers in Mortgage After Imputing : 0


Data.describe().T


# Now, we'll check for Target imbalancing before splitting:

sns.countplot('LoanOnCard',data=Data)

<AxesSubplot:xlabel='LoanOnCard', ylabel='count'>


Data.LoanOnCard.value_counts().to_frame()


# We can clearly see that the division of data between Loan: Yes or no is not very balanced.
# To balance data for splitting, we'll use Oversampeling method, 
# this will not let the important information get removed,compared to undersampling.

# We'll be performing train - test split:

# 1st well do splitting without balancing-

X=Data.drop(columns='LoanOnCard')
y=Data['LoanOnCard']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=10)


y_train.value_counts()

0.0    3153
1.0     333
Name: LoanOnCard, dtype: int64


y_test.value_counts()

0.0    1347
1.0     147
Name: LoanOnCard, dtype: int64


y.value_counts()

0.0    4500
1.0     480
Name: LoanOnCard, dtype: int64


from imblearn.over_sampling import SMOTENC
from imblearn.over_sampling import SMOTE

smote_nc=SMOTENC(categorical_features=[1,3,5],random_state=1) #specifying categorical column numbers
x_s,y_s=smote_nc.fit_resample(X,y)


y_s.value_counts()

1.0    4500
0.0    4500
Name: LoanOnCard, dtype: int64


# Again splitting the data on balanced sample
X_train, X_test, y_train, y_test = train_test_split(x_s, y_s, test_size=0.30, random_state=10)


# importing libraries for Logistic Reg and Gaussian model.
from sklearn.metrics import accuracy_score, confusion_matrix
from sklearn.linear_model import LogisticRegression
from sklearn.naive_bayes import GaussianNB
from sklearn.svm import SVC
from sklearn.metrics import classification_report
from sklearn import model_selection
import warnings

# Logistic regression model
logit = LogisticRegression()
logit.fit(X_train, y_train)
logit_pred = logit.predict(X_test)


logit.score(X_train,y_train)

0.8765079365079365


logit.score(X_test,y_test)

0.88


y_test.value_counts().sum()

2700


print('Report of our Logistic Regression model:\n',
      classification_report(y_test,logit_pred))

Report of our Logistic Regression model:
               precision    recall  f1-score   support

         0.0       0.89      0.87      0.88      1349
         1.0       0.87      0.89      0.88      1351

    accuracy                           0.88      2700
   macro avg       0.88      0.88      0.88      2700
weighted avg       0.88      0.88      0.88      2700


cm = confusion_matrix(y_test, logit_pred, labels=[0, 1])

d_cm = pd.DataFrame(cm, index = [i for i in ["Non-Loan Cust.","Loan Cust."]],
                  columns = [i for i in ["Non-Loan Cust.","Loan Cust."]])
plt.figure(figsize = (7,5))
sns.heatmap(d_cm, annot=True ,fmt='g', cmap = 'YlGnBu')
plt.show()


# naive bayes model:
X_train, X_test, y_train, y_test = train_test_split(x_s, y_s, test_size=0.30, random_state=5)

g_m = GaussianNB()
g_m.fit(X_train, y_train.ravel())
g_pred = g_m.predict(X_test)


# training data score
g_m.score(X_train, y_train)

0.8511111111111112


# testing data score
g_m.score(X_test, y_test)

0.8633333333333333


cm1 = confusion_matrix(y_test, g_pred, labels=[0, 1])

d1_cm = pd.DataFrame(cm1, index = [i for i in ["Non-Loan Cust.","Loan Cust."]],
                  columns = [i for i in ["Non-Loan Cust.","Loan Cust."]])
plt.figure(figsize = (7,5))
sns.heatmap(d1_cm, annot=True ,fmt='g', cmap = 'YlGnBu')
plt.show()


print('Report of our Naive Bayes model:\n',
      classification_report(y_test,g_pred))

Report of our Naive Bayes model:
               precision    recall  f1-score   support

         0.0       0.84      0.89      0.87      1341
         1.0       0.88      0.84      0.86      1359

    accuracy                           0.86      2700
   macro avg       0.86      0.86      0.86      2700
weighted avg       0.86      0.86      0.86      2700

	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree	Class
0	63.027818	22.552586	39.609117	40.475232	98.672917	-0.254400	Type_H
1	39.056951	10.060991	25.015378	28.995960	114.405425	4.564259	Type_H
2	68.832021	22.218482	50.092194	46.613539	105.985135	-3.530317	Type_H
3	69.297008	24.652878	44.311238	44.644130	101.868495	11.211523	Type_H
4	49.712859	9.652075	28.317406	40.060784	108.168725	7.918501	Type_H

	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree	Class
0	74.377678	32.053104	78.772013	42.324573	143.560690	56.125906	Type_S
1	89.680567	32.704435	83.130732	56.976132	129.955476	92.027277	Type_S
2	44.529051	9.433234	52.000000	35.095817	134.711772	29.106575	Type_S
3	77.690577	21.380645	64.429442	56.309932	114.818751	26.931841	Type_S
4	76.147212	21.936186	82.961502	54.211027	123.932010	10.431972	Type_S

	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree	Class
140	63.364339	20.024621	67.498705	43.339718	130.999258	37.556706	Type_S
141	67.513053	33.275590	96.283062	34.237463	145.601033	88.301486	Type_S
142	76.314028	41.933683	93.284863	34.380345	132.267285	101.218783	Type_S
143	73.635962	9.711318	63.000000	63.924644	98.727930	26.975787	Type_S
144	56.535051	14.377189	44.991547	42.157862	101.723334	25.773174	Type_S
145	80.111572	33.942432	85.101608	46.169139	125.593624	100.292107	Type_S
146	95.480229	46.550053	59.000000	48.930176	96.683903	77.283072	Type_S
147	74.094731	18.823727	76.032156	55.271004	128.405731	73.388216	Type_S
148	87.679087	20.365613	93.822416	67.313473	120.944829	76.730629	Type_S
149	48.259920	16.417462	36.329137	31.842457	94.882336	28.343799	Type_S

	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree	Class
244	60.044177	14.309656	58.038865	45.734521	105.131664	30.409133	Type_S
194	94.174822	15.380770	67.705721	78.794052	114.890113	53.255220	Type_S
268	95.382596	24.822631	95.157633	70.559965	89.307547	57.660841	Type_S
152	50.209670	29.760122	36.104007	20.449548	128.292515	5.740614	Type_H
29	56.447026	19.444499	43.577846	37.002527	139.189690	-1.859689	Normal
48	40.349296	10.194748	37.967747	30.154548	128.009927	0.458901	Normal
306	95.480229	46.550053	59.000000	48.930176	96.683903	77.283072	Type_S
137	35.703458	19.443253	20.700000	16.260205	137.540613	-0.263490	Type_H
262	118.144655	38.449501	50.838520	79.695154	81.024541	74.043767	Type_S
169	58.599529	-0.261499	51.500000	58.861028	102.042812	28.059697	Type_S

	count	mean	std	min	25%	50%	75%	max
P_incidence	310.0	60.496653	17.236520	26.147921	46.430294	58.691038	72.877696	129.834041
P_tilt	310.0	17.542822	10.008330	-6.554948	10.667069	16.357689	22.120395	49.431864
L_angle	310.0	51.930930	18.554064	14.000000	37.000000	49.562398	63.000000	125.742385
S_slope	310.0	42.953831	13.423102	13.366931	33.347122	42.404912	52.695888	121.429566
P_radius	310.0	117.920655	13.317377	70.082575	110.709196	118.268178	125.467674	163.071041
S_Degree	310.0	26.296694	37.559027	-11.058179	1.603727	11.767934	41.287352	418.543082

Project 2: Supervised Learning¶

This project comparises of 2 Parts:¶

Part 1: Healthcare:¶

1. Import and warehouse data:¶

Observation:¶

2. Data cleansing:¶

Observation:¶

Observation:¶

3. Data analysis & visualisation:¶

Observations:¶

4. Data pre-processing:¶

5. Model training, testing and tuning:¶

6. Conclusion and improvisation:¶

-----------END OF PROJECT 1---------------------END OF PROJECT 1---------------------END OF PROJECT 1----------¶

Part 2: Banking and Finance¶

PROJECT OBJECTIVE:¶

1. Import and warehouse data:¶

2. Data cleansing:¶

3. Data analysis & visualisation:¶

4. Data pre-processing:¶

5. Model training, testing and tuning:¶

In Conclusion, best fit model for our case is Logistic Regression model with Accuracy at 88%.¶

6. Conclusion and improvisation:¶

	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree	Class
0	38.505273	16.964297	35.112814	21.540976	127.632875	7.986683	Normal
1	54.920858	18.968430	51.601455	35.952428	125.846646	2.001642	Normal
2	44.362490	8.945435	46.902096	35.417055	129.220682	4.994195	Normal
3	48.318931	17.452121	48.000000	30.866809	128.980308	-0.910941	Normal
4	45.701789	10.659859	42.577846	35.041929	130.178314	-3.388910	Normal

	0
P_incidence	True
P_tilt	True
L_angle	True
S_slope	True
P_radius	True
S_Degree	True
Class	True

	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree	Class
215	129.834041	8.404475	48.384057	121.429566	107.690466	418.543082	Type_S
262	118.144655	38.449501	50.838520	79.695154	81.024541	74.043767	Type_S
263	115.923261	37.515436	76.800000	78.407825	104.698603	81.198927	Type_S

	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree	Class
183	81.10410	24.794168	77.887020	56.309932	151.839857	65.214616	Type_S
185	45.44375	9.906072	45.000000	35.537678	163.071041	20.315315	Type_S
280	37.90391	4.479099	24.710274	33.424811	157.848799	33.607027	Type_S

	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree	Class
242	85.290173	18.278890	100.744220	67.011283	110.660700	58.884948	Type_S
297	58.828379	37.577873	125.742385	21.250506	135.629418	117.314683	Type_S

	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree
0	-1.297382	-0.060842	-0.907111	-1.671791	0.727500	-0.564534
1	-0.318961	0.139676	-0.018377	-0.533124	0.593244	-0.762811
2	-0.948273	-0.863147	-0.271672	-0.575424	0.846843	-0.663672
3	-0.712457	-0.012034	-0.212495	-0.934945	0.828776	-0.859301
4	-0.868447	-0.691615	-0.504747	-0.605063	0.918821	-0.941394

	count	mean	std	min	25%	50%	75%	max
P_incidence	309.0	7.114050e-17	1.001622	-2.033920	-0.825260	-0.099700	0.737387	3.449381
P_tilt	309.0	9.651575e-17	1.001622	-2.413996	-0.688730	-0.115554	0.461181	3.187612
L_angle	309.0	2.705495e-16	1.001622	-2.045086	-0.805392	-0.116794	0.596001	3.977800
S_slope	309.0	-3.848055e-16	1.001622	-2.317633	-0.739479	-0.025779	0.778226	2.923046
P_radius	309.0	6.445761e-16	1.001622	-3.598089	-0.544462	0.029280	0.565699	3.391100
S_Degree	309.0	3.245336e-16	1.001622	-1.195467	-0.776291	-0.449361	0.525211	4.098903

	ID	Age	CustomerSince	HighestSpend	ZipCode	HiddenScore	MonthlyAverageSpend	Level
0	1	25	1	49	91107	4	1.6	1
1	2	45	19	34	90089	3	1.5	1
2	3	39	15	11	94720	1	1.0	1
3	4	35	9	100	94112	1	2.7	2
4	5	35	8	45	91330	4	1.0	2

ID	False
Age	False
CustomerSince	False
HighestSpend	False
ZipCode	False
HiddenScore	False
MonthlyAverageSpend	False
Level	False
Mortgage	False
Security	False
FixedDepositAccount	False
InternetBanking	False
CreditCard	False
LoanOnCard	False	True
0	4980	20

	count	mean	std	min	25%	50%	75%	max
ID	5000.0	2500.500000	1443.520003	1.0	1250.75	2500.5	3750.25	5000.0
Age	5000.0	45.338400	11.463166	23.0	35.00	45.0	55.00	67.0
CustomerSince	5000.0	20.104600	11.467954	-3.0	10.00	20.0	30.00	43.0
HighestSpend	5000.0	73.774200	46.033729	8.0	39.00	64.0	98.00	224.0
ZipCode	5000.0	93152.503000	2121.852197	9307.0	91911.00	93437.0	94608.00	96651.0
HiddenScore	5000.0	2.396400	1.147663	1.0	1.00	2.0	3.00	4.0
MonthlyAverageSpend	5000.0	1.937938	1.747659	0.0	0.70	1.5	2.50	10.0
Level	5000.0	1.881000	0.839869	1.0	1.00	2.0	3.00	3.0
Mortgage	5000.0	56.498800	101.713802	0.0	0.00	0.0	101.00	635.0
Security	5000.0	0.104400	0.305809	0.0	0.00	0.0	0.00	1.0
FixedDepositAccount	5000.0	0.060400	0.238250	0.0	0.00	0.0	0.00	1.0
InternetBanking	5000.0	0.596800	0.490589	0.0	0.00	1.0	1.00	1.0
CreditCard	5000.0	0.294000	0.455637	0.0	0.00	0.0	1.00	1.0
LoanOnCard	4980.0	0.096386	0.295149	0.0	0.00	0.0	0.00	1.0

	index	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree	Class
0	0	38.505273	16.964297	35.112814	21.540976	127.632875	7.986683	Normal
1	1	54.920858	18.968430	51.601455	35.952428	125.846646	2.001642	Normal
2	2	44.362490	8.945435	46.902096	35.417055	129.220682	4.994195	Normal
3	3	48.318931	17.452121	48.000000	30.866809	128.980308	-0.910941	Normal
4	4	45.701789	10.659859	42.577846	35.041929	130.178314	-3.388910	Normal
5	5	30.741938	13.354966	35.903526	17.386972	142.410107	-2.005373	Normal
6	6	50.913101	6.677000	30.896522	44.236102	118.151531	-1.057986	Normal
7	7	38.126589	6.557617	50.445075	31.568971	132.114805	6.338199	Normal
8	8	51.624672	15.969344	35.000000	35.655328	129.385308	1.009228	Normal
9	9	64.311867	26.328369	50.958964	37.983498	106.177751	3.118221	Normal
10	10	44.489275	21.786433	31.474154	22.702842	113.778494	-0.284129	Normal
11	11	54.950970	5.865353	53.000000	49.085617	126.970328	-0.631603	Normal
12	12	56.103774	13.106307	62.637020	42.997467	116.228503	31.172767	Normal
13	13	69.398818	18.898407	75.966361	50.500411	103.582540	-0.443661	Normal
14	14	89.834676	22.639217	90.563461	67.195460	100.501192	3.040973	Normal
15	15	59.726140	7.724873	55.343485	52.001268	125.174221	3.235159	Normal
16	16	63.959522	16.060945	63.123736	47.898577	142.360125	6.298971	Normal
17	17	61.540599	19.676957	52.892229	41.863642	118.686268	4.815031	Normal
18	18	38.046551	8.301669	26.236830	29.744881	123.803413	3.885773	Normal
19	19	43.436451	10.095743	36.032224	33.340707	137.439694	-3.114451	Normal
20	20	65.611802	23.137919	62.582179	42.473883	124.128001	-4.083298	Normal
21	21	53.911054	12.939318	39.000000	40.971736	118.193035	5.074353	Normal
22	22	43.117951	13.815744	40.347388	29.302207	128.517722	0.970926	Normal
23	23	40.683229	9.148437	31.021593	31.534792	139.118472	-2.511619	Normal
24	24	37.731992	9.386298	42.000000	28.345694	135.740926	13.683047	Normal
25	25	63.929470	19.971097	40.177050	43.958373	113.065939	-11.058179	Normal
26	26	61.821627	13.597105	64.000000	48.224523	121.779803	1.296191	Normal
27	27	62.140805	13.960975	58.000000	48.179830	133.281834	4.955106	Normal
28	28	69.004913	13.291790	55.570143	55.713123	126.611622	10.832011	Normal
29	29	56.447026	19.444499	43.577846	37.002527	139.189690	-1.859689	Normal
30	30	41.646916	8.835549	36.031975	32.811367	116.555168	-6.054538	Normal
31	31	51.529358	13.517847	35.000000	38.011510	126.718516	13.928331	Normal
32	32	39.087264	5.536602	26.932038	33.550662	131.584420	-0.759461	Normal
33	33	34.649922	7.514783	43.000000	27.135140	123.987741	-4.082938	Normal
34	34	63.026300	27.336240	51.605017	35.690060	114.506608	7.439870	Normal
35	35	47.805559	10.688698	54.000000	37.116861	125.391138	-0.402523	Normal
36	36	46.637864	15.853717	40.000000	30.784147	119.377603	9.064582	Normal
37	37	49.828135	16.736435	28.000000	33.091700	121.435558	1.913307	Normal
38	38	47.319648	8.573680	35.560252	38.745967	120.576972	1.630664	Normal
39	39	50.753290	20.235060	37.000000	30.518231	122.343516	2.288488	Normal

	P_incidence	P_tilt	L_angle	S_slope	P_radius	S_Degree	Class
0	38.505273	16.964297	35.112814	21.540976	127.632875	7.986683	Normal
1	54.920858	18.968430	51.601455	35.952428	125.846646	2.001642	Normal
2	44.362490	8.945435	46.902096	35.417055	129.220682	4.994195	Normal
3	48.318931	17.452121	48.000000	30.866809	128.980308	-0.910941	Normal
4	45.701789	10.659859	42.577846	35.041929	130.178314	-3.388910	Normal
5	30.741938	13.354966	35.903526	17.386972	142.410107	-2.005373	Normal
6	50.913101	6.677000	30.896522	44.236102	118.151531	-1.057986	Normal
7	38.126589	6.557617	50.445075	31.568971	132.114805	6.338199	Normal
8	51.624672	15.969344	35.000000	35.655328	129.385308	1.009228	Normal
9	64.311867	26.328369	50.958964	37.983498	106.177751	3.118221	Normal
10	44.489275	21.786433	31.474154	22.702842	113.778494	-0.284129	Normal
11	54.950970	5.865353	53.000000	49.085617	126.970328	-0.631603	Normal
12	56.103774	13.106307	62.637020	42.997467	116.228503	31.172767	Normal
13	69.398818	18.898407	75.966361	50.500411	103.582540	-0.443661	Normal
14	89.834676	22.639217	90.563461	67.195460	100.501192	3.040973	Normal
15	59.726140	7.724873	55.343485	52.001268	125.174221	3.235159	Normal
16	63.959522	16.060945	63.123736	47.898577	142.360125	6.298971	Normal
17	61.540599	19.676957	52.892229	41.863642	118.686268	4.815031	Normal
18	38.046551	8.301669	26.236830	29.744881	123.803413	3.885773	Normal
19	43.436451	10.095743	36.032224	33.340707	137.439694	-3.114451	Normal
20	65.611802	23.137919	62.582179	42.473883	124.128001	-4.083298	Normal
21	53.911054	12.939318	39.000000	40.971736	118.193035	5.074353	Normal
22	43.117951	13.815744	40.347388	29.302207	128.517722	0.970926	Normal
23	40.683229	9.148437	31.021593	31.534792	139.118472	-2.511619	Normal
24	37.731992	9.386298	42.000000	28.345694	135.740926	13.683047	Normal
25	63.929470	19.971097	40.177050	43.958373	113.065939	-11.058179	Normal
26	61.821627	13.597105	64.000000	48.224523	121.779803	1.296191	Normal
27	62.140805	13.960975	58.000000	48.179830	133.281834	4.955106	Normal
28	69.004913	13.291790	55.570143	55.713123	126.611622	10.832011	Normal
29	56.447026	19.444499	43.577846	37.002527	139.189690	-1.859689	Normal
30	41.646916	8.835549	36.031975	32.811367	116.555168	-6.054538	Normal
31	51.529358	13.517847	35.000000	38.011510	126.718516	13.928331	Normal
32	39.087264	5.536602	26.932038	33.550662	131.584420	-0.759461	Normal
33	34.649922	7.514783	43.000000	27.135140	123.987741	-4.082938	Normal
34	63.026300	27.336240	51.605017	35.690060	114.506608	7.439870	Normal
35	47.805559	10.688698	54.000000	37.116861	125.391138	-0.402523	Normal
36	46.637864	15.853717	40.000000	30.784147	119.377603	9.064582	Normal
37	49.828135	16.736435	28.000000	33.091700	121.435558	1.913307	Normal
38	47.319648	8.573680	35.560252	38.745967	120.576972	1.630664	Normal
39	50.753290	20.235060	37.000000	30.518231	122.343516	2.288488	Normal

	HiddenScore
1	1472
2	1296
4	1222
3	1010

	Level
1	2096
3	1501
2	1403

	count	mean	std	min	25%	50%	75%	max
ID	4980.0	2510.345382	1438.011129	10.0	1265.75	2510.5	3755.250	5000.0
Age	4980.0	45.352610	11.464212	23.0	35.00	45.0	55.000	67.0
CustomerSince	4980.0	20.117671	11.468716	-3.0	10.00	20.0	30.000	43.0
HighestSpend	4980.0	73.852410	46.070090	8.0	39.00	64.0	98.000	224.0
ZipCode	4980.0	93152.420482	2123.660073	9307.0	91911.00	93407.0	94608.000	96651.0
MonthlyAverageSpend	4980.0	1.939536	1.750006	0.0	0.70	1.5	2.525	10.0
Mortgage	4980.0	56.589759	101.836758	0.0	0.00	0.0	101.000	635.0

	count	mean	std	min	25%	50%	75%	max
HighestSpend	4980.0	71.468474	42.834544	8.0	39.0	64.0	93.0	185.0
MonthlyAverageSpend	4980.0	1.559255	1.160226	0.0	0.7	1.3	2.2	5.2
Mortgage	4980.0	38.027510	66.111473	0.0	0.0	0.0	38.0	252.0